When should I use UltraWarm nodes?

UltraWarm nodes are ideal for storing data that you query infrequently but need to keep accessible — typically older log data or time-series data beyond the hot retention period. UltraWarm uses S3-backed storage at roughly one-tenth the cost of hot EBS-attached storage. Data on UltraWarm nodes is read-only. A common pattern is to keep 7-14 days of data on hot nodes for active querying and move older data to UltraWarm for up to several months of retention at lower cost.

Do I need dedicated master nodes?

AWS strongly recommends dedicated master nodes for production domains. Master nodes manage cluster state, shard allocation, and index operations — without dedicated masters, data nodes handle these tasks, which can impact query performance during cluster events. Use three master nodes (for quorum) in production. For domains with fewer than 10 data nodes and non-critical workloads, you can skip dedicated masters, but expect reduced stability during blue/green deployments or node failures.

OpenSearch Cost Estimator

StorageAWS

Estimate Amazon OpenSearch domain costs with data, master, and warm nodes.

Last verified: May 2026

OpenSearch Configuration

Data Nodes

Instance Type

EBS Storage per Node (GB)

gp3 EBS at $0.08/GB/month per data node.

Dedicated Master Nodes

Raw Output

Output will appear here...

See It in Action

Your team's log analytics OpenSearch domain has 12 r5.4xlarge.search data nodes ($7,200/month) holding 90 days of logs. The estimator helps you redesign: 6 r6g.4xlarge.search hot nodes (Graviton, 20% cheaper) holding 14 days, plus 6 ultrawarm1.large.search nodes holding the older 76 days at 1/5 the cost. New monthly cost: $4,100. Annual savings: $37K, with zero impact on query latency for the 95% of queries hitting recent data.

What This Tool Does

Amazon OpenSearch Service (successor to Amazon Elasticsearch Service) provides managed search and analytics domains with support for OpenSearch and legacy Elasticsearch engines. Domain costs depend on the number and type of data nodes, optional dedicated master nodes for cluster stability, warm nodes using UltraWarm for cost-effective storage of infrequently queried data, and EBS storage volumes. Additional costs include data transfer, optional features like fine-grained access control, and cross-cluster search. This estimator models monthly costs for production-grade OpenSearch deployments across different node configurations.

Technical Details

The estimator computes OpenSearch domain cost as: data node count × instance hourly rate × hours, plus optional master node count × rate, plus optional UltraWarm node count × rate, plus EBS storage × per-GB-month rate, plus data transfer (intra-region free, cross-AZ for Multi-AZ). Output presents per-component breakdown plus comparison against self-managed Elasticsearch on EC2 (showing both the operational savings and the cost delta).

Common Use Cases

1Sizing an OpenSearch domain with the right balance of data nodes, master nodes, and warm nodes for log analytics
2Estimating costs for migrating from self-managed Elasticsearch to Amazon OpenSearch Service
3Comparing costs between different instance families (general purpose, compute-optimized, memory-optimized) for search workloads

Common Questions

When should I use UltraWarm nodes?: UltraWarm nodes are ideal for storing data that you query infrequently but need to keep accessible — typically older log data or time-series data beyond the hot retention period. UltraWarm uses S3-backed storage at roughly one-tenth the cost of hot EBS-attached storage. Data on UltraWarm nodes is read-only. A common pattern is to keep 7-14 days of data on hot nodes for active querying and move older data to UltraWarm for up to several months of retention at lower cost.
Do I need dedicated master nodes?: AWS strongly recommends dedicated master nodes for production domains. Master nodes manage cluster state, shard allocation, and index operations — without dedicated masters, data nodes handle these tasks, which can impact query performance during cluster events. Use three master nodes (for quorum) in production. For domains with fewer than 10 data nodes and non-critical workloads, you can skip dedicated masters, but expect reduced stability during blue/green deployments or node failures.

Expert Tips

TIP

UltraWarm is the killer cost optimization for log analytics workloads. Hot storage costs ~$0.135/GB/month; UltraWarm costs ~$0.024/GB/month — an 82% reduction. Move data older than 7-14 days to UltraWarm via Index State Management policies. Most query patterns only hit the last few days of data anyway.

TIP

Always provision dedicated master nodes for production. The cost is small (3x m6g.large.search = ~$200/month) but the stability benefit is massive. Without dedicated masters, data nodes do double duty during cluster events (rebalancing, blue/green deploys) — query latency tanks during those windows.

TIP

Graviton-based instances (r6g, m6g) for OpenSearch deliver 20-30% better price-performance than x86 equivalents. Default to ARM unless you have specific x86 dependencies (which is rare for OpenSearch since it's all JVM). The migration is a single instance type change in the domain config.

Related Learning Guides

Cost Optimization Strategies26 min read

Was this tool helpful?

Disclaimer: This tool runs entirely in your browser. No data is sent to our servers. Always verify outputs before using them in production. AWS, Azure, and GCP are trademarks of their respective owners.